使用 SF 聚合栅格中的值

Question

使用 SF 聚合栅格中的值

我需要的是聚合每个栅格的某些指标的值。假设我们有一些数据 - 坐标和值，我想创建一个热图。

首先，我使用简单的要素框架创建了一个网格和栅格。现在我需要获取 df 中的每个坐标并检查它们是否在一个光栅中。然后为每个栅格计算任何其他聚合函数的平均值。


# Packages ----------------------------------------------------------------


library(raster)
library(tidyverse)
library(sf)
library(sp)


# Border Data and Grid ----------------------------------------------------

Regions <- getData("GADM", country= "CZE", level = 1)
Regions %>% 


Regions <-
  Regions %>% st_as_sf()

grid_spacing <- 0.25

polygony <- st_make_grid(Regions, square = T,
                         cellsize = c(grid_spacing, grid_spacing)) %>%
  st_sf()

plot(polygony, col = 'white')
plot(st_geometry(CZ), add = T)

A = st_intersection(polygony, CZ)


# Artifial Values to be use -----------------------------------------------
df <- 
  tibble(long = runif(500, 13.27857, 14),
         lat = runif(500, 49, 50),
         price = rnbinom(500, size = 40, 0.3))

df %>% 
  ggplot(aes(long, lat, color = values)) + 
  geom_point()

A <- 
  A %>% 
  # Here Val shall be calculated as a mean of observations within each rastel grid cell
  mutate(val = rnorm(n = 202))

# Since not every raster cell has observations inside some NAN will be present
A[5, 'val']  = NaN
A[25, 'val']  = NaN

A %>% 
  ggplot(aes(fill = val)) + 
  geom_sf()



# What I need -------------------------------------------------------------

# For each raster in A calculate the mean of price for coordinates within this raster grid.

我基本上需要 sf 框架内的 https://www.rdocumentation.org/packages/raster/versions/3.4-5/topics/rasterize 等价物。

但是，我想绘制成示例中的样子，我确实需要栅格具有给定网格的形状。

Answer 1

反正你用的是光栅包

library(raster)
cze <- getData("GADM", country= "CZE", level = 1)
r <- raster(cze, res=0.25)
df <- data.frame(long = runif(500, 13.27857, 14),
         lat = runif(500, 49, 50),
         price = rnbinom(500, size = 40, 0.3))

r <- rasterize(df[,c("long", "lat")], r, df$price, mean, background=0)
p <- as(r, "SpatialPolygonsDataFrame")
p <- crop(p, cze)

library(sf)
s <- st_as_sf(p)
plot(p)

Answer 2

要聚合网格中的点值并留在 sf 中，您可以执行以下操作：

library(dplyr)
library(sf)

# get data. using your example, we'll take spatial data from the raster package
Regions <- raster::getData("GADM", country= "CZE", level = 1)

# convert spdf to sf
Regions <- Regions %>% st_as_sf()

# create a grid sf object
grid_spacing <- 0.25

polygony <- st_make_grid(Regions, square = T,
                         cellsize = c(grid_spacing, grid_spacing)) %>%
  st_sf() %>% 
  mutate(ID = row_number()) # add a unique ID to each grid cell

# clip grid to shape of country polygons
A <- st_intersect(polygony, Regions)

# create fake data with coordinates and prices
df <- tibble(long = runif(500, 13.27857, 14),
         lat = runif(500, 49, 50),
         price = rnbinom(500, size = 40, 0.3))

# convert the df to sf point layer
points <- st_as_sf(df, coords = c("long", "lat"), crs = st_crs(A))

# spatially join grid to points, so that each point is assigned the grid ID into which it falls
pointsID <- st_join(points, A)

# group and summarize point values by grid ID
pointsID <- pointsID %>% 
  as.data.frame() %>% 
  group_by(ID) %>% 
  summarize(avg_price = mean(price))

# join aggregated values back to your grid
A<- left_join(A, pointsID, by = "ID")

plot(A["avg_price"])

使用 SF 聚合栅格中的值

Aggregate values in raster using SF

maps

r

polygon

raster

sf